Everything was working just fine when suddenly the framework stopped working.

You may have exceeded the license point limit allowed for your system. This can happen if the AutoTagAnalyticPoint property on the AnalyticService property sheet is set to true. This property should be set to false, unless you are configuring the system for the first time.

My chart is not displaying data correctly.

Confirm that the number of bindings is correct. If more bindings are specified than are required, nothing happens. Binding support is as follows:

I configured the Time Range and Interval, but my Spectrum Chart is completely blank.

If the Spectrum Chart binding returns less than three data points (that is 0, 1 or 2 data points), the chart fails and reports an exception. To investigate, calculate the total number of expected data points by dividing Time Range by Interval. and adjust the Time Range or Interval to report three or more data points.

How do I limit the data source to average for a min/max of one day in an algorithm?

Enable User Request Rollup on the data source and set Rollup to min or max as required.

I can add an Analytic Web Chart Binding to a web widget and configure it, but I cannot seem to get it to work.

After adding the Analytic Web Chart Binding, double-click the widget to open the Properties view, and delete the default WbView Binding.

I tried to use the Analytic Value Binding, but where do I define the time range and rollup functions?

The Analytic Value Binding gives the current value without any rollup.

I imported histories and points under my NiagaraNetwork from my remote host and tagged them, but I cannot visualize anything.

Enable the Persist Fetched Tags on the AX Property Sheet of the NiagaraNetwork driver, right-click on the Niagara Driver and click Actions > Force Update Niagara Proxy Points. This should apply a direct n:history tag to each NiagaraNetwork point where the point from the remote station has a history that has been imported to this station.

I want my algorithm to return trend results if a value was less than or equal to zero. Algorithm results show in a bound label but not on a web chart.

The bound label is likely resolving an analytic value request, whereas the web chart is resolving an analytic trend request. It it likely the control point does not have an n:history tag used by the framework to locate the history data. For a NiagaraNetwork, right-click the Config > Drivers > NiagaraNetwork folder in the Nav tree and click Actions > Force Update Niagara Proxy Points. This applies a direct n:history tag to any point that lacks a history extension.

For all other networks, use a Program Object in the V2 bog file to add an n:history tag to each point.

I’m running in a JACE-8000. My Web chart causes a server session time-out.

Check to see if you have specified a COV (change of value) point directly in the binding. A COV point that changes frequently can cause CPU spikes. As a best practice, instead of specifying the COV point directly, specify its parent in the binding. For other point types, configure a less frequent Refresh Rate in the binding to minimize CPU spikes.

I notice that, when I run PX views, my JACE-8000 slows down, and sometimes reports server session time out errors.

For best performance on the JACE-8000, limit the number of points configured in a PX view to 100–200 with no more than 200 tags, 500 history rollups and five bindings. For more complicated configurations, set up PX graphics in a Supervisor station running on a PC.

For an analytic request (binding, alert, etc.), the unit of measure output from an algorithm is not being converted correctly or it does not match the unit set in the algorithm.

Check the algorithm to ensure that the correct unit of measure is defined (by facets on the algorithm’s property sheet). Algorithms perform no unit conversion from data source to algorithm output. The unit of measure defined in the algorithm’s facets is directly output with the calculated value. This makes it imperative to define the correct unit of measure on the algorithm’s property sheet.

For a series of chained algorithms, for example: Algorithm 1 becomes a data source for Algorithm 2, which, in turn, becomes a data source for a Px binding, the system converts the output unit from Algorithm 1 (assuming the Algorithm 1 unit is defined in its facets) to the unit specified for the data source in Algorithm 2. If Algorithm 1 has no unit defined, and Algorithm 2 has the unit defined, Algorithm 2 applies its unit of measure to the input it receives from Algorithm 1.

I configured a data source for my web chart, but the system says that the data source is not available.

Check the station log to identify the origin of the request for data.

Figure 76.   Station log

A request from a graphic is classified as a user request. This is followed by which point, node, and user are involved in this request, as well as the name of the algorithm in which the request was made. The multiple origins in the screen capture example represent nested algorithms.

If this does not help you solve the problem, open the Property Sheet for the AnalyticService, and set Skip Data Source Cache to true. The framework engine caches memory to improve response time. Disabling memory cache, by setting this property to true, causes the system to display the current error in the station log.

I am trying to figure out why an alert occurred.

Check the station log. It shows which point generated the user request that triggered the alert along with the alert name.

I tagged points to be used with analytics and my AnalyticService is now in fault and all analytic requests fail. What happened?

There is more than one reason for this to happen:

I set up a Source Name for an alert expecting that the text and BFormat I entered would display in the alarm console. Instead, the alarm console displays the default BFormat (%node.navName%_%alert.name%) in its Source column.

You have a syntax error in your BFormat. Clicking the Notes button displays this message, “The BFormat value for sourceName is invalid for alert AnalyticAlert.” where AnalyticAlert is replaced by the name you configured for the alert. Check the Reference manual for examples of BFormat syntax.

After changing an algorithm, refreshing cache can take as long as 20 minutes. What is going on?

Refresh Cache calculates data memory requirements again by searching all hierarchies. Depending on the hierarchical structure of your data, this could take some time.